An look at how the New York Times uses slang in crossword puzzles and what it reveals about the publication's culture.
For decades the people making decisions about what should be in the New York Times crossword have been straight white men. According to some crossword constructors, the topic of who the crossword is written for has been an active point of discussion and debate in crossword circles.
I thought studying how slang appears in the crossword could contribute to this discussion of the crossword’s intended audience.
Slang is a transitory language developed by diverse subgroups of society. Before social media and online communication, it took time for this language to enter the main culture's vocabulary. And it’s unusual for slang to remain in use for more than a few years - what is cool for one generation is old-fashioned for the next. Dated slang often drifts into obsolescence or becomes accepted into the standard language. A study of slang, when it enters the puzzle, how it enters the puzzle, who has control over how it gets defined (clued), and how this is changing over time could help shed light on the crossword culture and its actors.
For this piece, I analyzed 46 years of crossword answers and clues, primarily looking at words that are clued with the modifiers: "slang," "colloquially," "modern lingo," "informally," and "in textspeak.”
I wrote a Python method that scrapes the data using BeautifulSoup from XWordInfo.com for the years 1976-2022. I chose these years because I wanted to examine how the puzzle changed after Shortz assumed his position as editor but I didn't think it was necessary to collect it all the way from 1940's.
The data analysis was done using the Python library Pandas. Some of the graphs were created using Altair. Crosswords.ipynb was my prelim analysis (this notebook only analyzes a subset of the whole data)- most of the analysis isn't focused on the story that I wrote, rather it was the path to the analysis I did for my story - but this notebook is more interesting than the other ones. "More Crossword Analysis.ipynb" was what I used for the story: I expanded the dataset to include all the years till 2022 and used it to create .csv's to create visuals on Datawrapper.
Graph1, Graph2 and Graph3 are the visuals in the order as they are seen in the report. They were created in Datawrapper, edited in Adobe Illustrator and uploaded to html using ai2html.js.
“Crossword Editor Opens the Door to Innovation across the Board Shortz Story.” Baltimore Sun, https://www.baltimoresun.com/news/bs-xpm-1993-11-22-1993326127-story.html.
“In the CROSS-FIRE.” Chicago Tribune, https://www.chicagotribune.com/news/ct-xpm-1993-11-07-9311080002-story.html.
Aronow, Isaac. “How Word Lists Help - or Hurt - Crossword Puzzles.” The New York Times, The New York Times, 6 Oct. 2021, https://www.nytimes.com/2021/10/06/crosswords/wordlists-for-constructing-puzzles.html.
Zhou, Yanchun, and Yanhong Fan. “Home-Academy Publication.” Home-Academy Publication, Changchun University of Science and Technology, Changchun, China, https://academypublication.com/.
“Unpack.” GRAMMARIST, 16 Apr. 2022, https://grammarist.com/new-words/unpack/.
Apart from the sources that I have listed, I messaged a few NYT Crossword constructors on Twitter who were incredibly kind with their time: Aimee Lucido, Will Nediger and Robyn Weintraub. While Aimee and Will are featured in the story, Robyn's comments were about social media and word lists -> an angle I didn't have too much time to research and explore for this piece but a important topic to address in the future